Background/Objectives: Glioblastoma (GBM) is the most lethal primary brain malignancy in adults, with a median overall survival of approximately 15 months. Temozolomide (TMZ) resistance develops in virtually all patients, and no second-line regimen has improved outcomes over the past two decades. The DNA methyltransferase inhibitor decitabine (DAC) has attracted interest as a chemosensitizer, but whether it directly reverses the TMZ-resistance transcriptome or operates through distinct, complementary mechanisms has not been tested at multi-omics resolution. Methods: We performed an integrative six-layer multi-omics analysis across five public GEO datasets (bulk RNA-seq, EPIC 850K methylation, and 21,676 single cells) re-purposed from studies conducted for unrelated aims, formally tested DAC-mediated reversal of the TMZ-resistance transcriptome across 11,707 genes, mapped pharmacogenomic targets with DGIdb v5, and built an exploratory, hypothesis-generating 11-gene prognostic model internally validated in TCGA-GBM (n = 166) and externally tested in the independent CPTAC-GBM cohort (n = 96). Results: DAC reprogrammed transcription across 1114–1882 differentially expressed genes per cohort and reactivated 146 direct epigenetic targets, identifying INPP5D/SHIP1 as the top-ranked direct epigenetic-reactivation target. Genome-wide reversal analysis across 11,707 co-detected genes showed a negligible effect (Spearman ρ = 0.073), but single-cell analysis revealed significant per-cell attenuation of MES-like and stem-like programs (Δ = −0.071 and −0.135, respectively; both p < 0.001). The 11-gene risk model achieved a Harrell’s C-index of 0.706 (apparent); after correcting for the two-stage gene selection with a full-pipeline bootstrap, the optimism-corrected C-index was 0.63, and external validation in an independent cohort (CPTAC-GBM, n = 96) showed only near-chance discrimination (C-index 0.55), indicating that the signature does not generalize and is exploratory. Pharmacogenomic mapping yielded 734 unique therapeutic agents (230 FDA-approved) across 69 druggable targets after excluding AR. Most of these agents are not GBM-directed, so this catalog-level mapping is hypothesis-generating rather than a set of therapeutic recommendations. Conclusions: DAC does not broadly reverse the TMZ-resistant transcriptome but acts through three complementary mechanisms: epigenetic reactivation of INPP5D/SHIP1, cancer-testis-antigen and type I interferon induction, and per-cell attenuation of mesenchymal–stem-like transcriptional intensity, supporting hypotheses for rationally designed DAC-based combination therapy in TMZ-resistant GBM.

PTEN loss activates PI3K–AKT signaling, and mTORC1 stimulates pyrimidine synthesis through S6K1-mediated CAD phosphorylation. This exploratory public-data analysis tested whether PTEN deletion in primary prostate cancer is associated with CAD, mTORC1 and pyrimidine-synthesis expression, and whether these survive protein-level and confounder analyses. In 485 primary tumors from The Cancer Genome Atlas Prostate Adenocarcinoma (TCGA-PRAD; 159 PTEN-deleted), deletion was associated with higher CAD mRNA (Hedges’ g = 0.57) and higher Hallmark mTORC1 scores (g = 0.54). In 342 tumors with reverse-phase protein-array data, deletion was accompanied by lower PTEN protein and higher AKT Ser473 phosphorylation but not by a higher pre-specified mTORC1 phospho-output composite (g = −0.02), itself uncorrelated with the transcriptional score. In a pre-specified model adding purity, TP53, RB1, ERG fusion and batch (n = 276), the PTEN–CAD association was absent (β = −0.002; p = 0.96), whereas the mTORC1-score association persisted. In a post hoc decomposition, ERG fusion, which co-occurs with PTEN deletion (odds ratio 3.6), changed the coefficient most. Across eight cohorts, a small PTEN mRNA–CAD association (g = 0.23) was attenuated to 0.14 after ERG adjustment, with overlapping confidence intervals (Hartung–Knapp p = 0.071). Bulk-tumor pathway signatures and PTEN–CAD associations need protein-level and ERG-aware interpretation.